OneGoodArea 1.0, the deterministic data and intelligence layer for UK area workflows. This is the buyer-facing record of what shipped, from launch forward.
The 1.0 launch, open by default. Click to collapse. Each entry tagged feature, improvement, or fix.
The deterministic data and intelligence layer for UK area workflows, live. Four products over one signal store and one frozen scoring engine.
Per-area values across seven categories (crime, deprivation, property, schools, amenities, transport, environment). Each carries its source, observed period, confidence, and national and regional percentile.
Deterministic 0-100 composites over five weighted dimensions per preset. No AI in the scoring path: the same input and engine version produce the same number every time. Every response is versioned and pinnable via the X-Engine-Version header for model risk registers.
Track portfolios of areas and detect material change over the time-series store. Fires signal.changed webhooks, with sample-size gating to suppress small-sample noise.
A typed query plane. Send a programmatic plan or a natural-language question; both run through the same deterministic executor and echo the executed plan for replay. Peers (k-NN), insights (peer-relative anomalies), and forecast (linear projection). The AI plans the query; it never sets the numbers.
Per-org signal bundles, saved scoring presets, methodology pinning, peer cohorts, three-tier RBAC, white-label, and per-key IP allowlist.
Tell us what you’d like to see next. Most items on this page started as an email.